Good one! Glad you got it resolved.

I took the liberty of adding this tip to the ThingsToKnow doc page.

Btw: Did you also happen to get “kernel: ip_conntrack: table full, dropping 
packet.” in the syslog while this happened?

From: [email protected] [mailto:[email protected]] On 
Behalf Of Samuel
Sent: Wednesday, November 14, 2012 5:35 PM
To: uWSGI developers and users list
Subject: Re: [uWSGI] why some workers are much more busy than the others

Hi guys,

I resolved this problem, and I'm very happy to let you know how to resolve it:

up ip_conntrack_max = 65560 to 6556000

that is. everything done.

If anyone met such same problem, please check this number before doing anything 
else.



On Wed, Nov 14, 2012 at 11:01 AM, Jeff Van Voorst 
<[email protected]<mailto:[email protected]>> wrote:
I don't know if ab is distributed and how it sends requests.  You will have to 
try it and see.

Yes bees is tied to ec2 instances, but someone could probably modify it to come 
from a compute cluster or other network of computers (I haven't looked at the 
code).


--Jeff
_______________________________________________
uWSGI mailing list
[email protected]<mailto:[email protected]>
http://lists.unbit.it/cgi-bin/mailman/listinfo/uwsgi



--
吴焱红(Samuel)

博客: blog.shanbay.com<http://blog.shanbay.com>


_______________________________________________
uWSGI mailing list
[email protected]
http://lists.unbit.it/cgi-bin/mailman/listinfo/uwsgi

Reply via email to